Guo C, Jeong HH, Hsieh YC, Klein HU, Bennett DA, De Jager PL, Liu Z, Shulman JM.
Tau Activates Transposable Elements in Alzheimer's Disease.
Cell Rep. 2018 Jun 5;23(10):2874-2880.
PubMed.
I first heard of the idea of transposon mobilization in human brain disease from Joshua Dubnau, then at Cold Spring Harbor. Josh and his colleagues published a paper providing evidence for activation of transposable elements in association with TDP pathology in ALS and FTLD (Li et al., 2012). Here, Guo et al. use a large RNA-seq dataset to show correlations between tau pathology and activation of transcription from transposable element loci in human brain, and in flies transgenic for human tau. This is an intriguing paper, reinforcing an idea which sounds less novel now than it did in 2012.
For a tauist, there is a "black box" to this story. What is it about tau that might trigger such a catastrophic response in a neuron? The human brain work correlates transposon activation with neurofibrillary tangle formation, but in Drosophila such aggregation of over expressed human tau is rare, and cell death usually occurs without tau filament formation. It is also far from clear how transcription from these loci will relate to cell death. Is this a sign of a sick cell - or a critical step in tau-mediated cell death? The complexity of the transcriptional activation will make this difficult to dissect.
Almost half of our DNA content consists of ancient virus like elements called retrotransposons (RTEs). RTEs are capable of copying themselves and then reinserting their new copies into our chromosomes. Our genomes invest heavily in keeping this half of our DNA quiescent because the wholesale activation of RTEs is incredibly damaging to the cell in a variety of ways.
There is now accumulating evidence to support the hypothesis that RTEs may contribute to the cellular toxicity that causes a variety of age-dependent neurodegenerative disorders, including amyotrophic lateral sclerosis, frontotemporal dementia, and macular degeneration. This study adds evidence that RTEs may be at the heart of AD.
It is important to recognize that although RTEs make up a vast fraction of our DNA, they have been almost completely ignored by human genetics (and by most geneticists studying animal models) until recently. The reason for this is that the standard computational algorithms for examining troves of genomic data begin by throwing out almost all the RTE sequences before the analysis really gets going. That has to stop because evidence is accumulating that RTEs may contribute to many age-related diseases, including cancer and neurodegeneration, and even normal aging. So it now becomes quite important to stop ignoring RTE sequences. By focusing on RTE sequences, these authors have, in fact, found strong evidence that many RTEs are highly expressed in AD brains and in a Drosophila model of tau pathology. It now becomes important to ask if such expression is a cause of or a consequence of the disease state.
This study investigated whether transposable elements (TEs) are differentially expressed in association with tau pathology in Alzheimer's disease (AD), using human post-mortem brain samples and a Drosophila model of AD. With regards to the human analysis, the cohort size (n=636) is impressive, and the observation of association with neurofibrillary tangles is intriguing. I also very much appreciated the circumspect discussion.
At the same time, I have a lot of questions. For example, it is not clear to me how big an absolute expression change was required for a given TE group to be elevated and considered significantly associated with a given tau pathology parameter in the linear regression statistics. 2%, 20% or 200%? As the authors note, the only autonomous human TE (L1Hs) was not at all associated with tangles, making it unclear how TE activity would contribute to genomic instability. The authors propose instead that some TE families are activated by chromatin relaxation promoted by tau pathology, which has been reported previously, and this makes more sense.
It is also interesting that the 5' end of one of the significant L1 subfamily hits (e.g. L1MB4_5 - copy number I think < 100) is significantly associated with tangles, whilst the 3' end of the same subfamily (L1MB4 - copy number ~9,200) is not. Why is that and how does it fit with the chromatin relaxation model? It would be useful to know the copy numbers of all of the human TE families with significant associations in the study.
Finally, I would be interested to see if these batch results replicate if one simply aligns the RNA-seq directly to the genome and intersects those data with the genomic coordinates of each individual TE; all except for the youngest TE subfamilies are highly mappable with the length and quality of RNA-seq reads available here, and this could generate some interesting examples of individual TE copies that are associated with pathology. In sum, the study is preliminary but very interesting and leads to many additional questions.
This paper by Guo and colleagues on the activation of transposable elements, and, in particular, of the LTR of HERV-Fc1 endogenous retrovirus in tau-associated lesions of human brains from patients who died with Alzheimer’s disease is interesting, as AD prevalence is increasing in elderly. Tau mechanisms of action were not specified in the paper, but a recent review notes that “although tau protein is well-known for its key role in stabilizing and organization of axonal microtubule network, it bears a broad range of functions including DNA protection and participation in signalling pathways,” and that tau can be modified by a variety of cellular enzymes, which in turn broaden tau’s function and interaction spectrum (Borna et al., 2018). This means that, in tauopathies, multiple mechanisms can be modified at different levels in DNA conformation and expression, and that it is hard to discriminate which effect is disease-related, and which is not.
In this study the only statistically significant alteration detected for transposable elements was in the transcriptional signature of HERV-Fc1 LTR. This could be relevant to correlate HERV-Fc1 to neurodegeneration, pending that transcripts driven by the LTR promoter are associated with some potentially neuropathogenic HERV-Fc1 ORFs, especially products of the env gene. It is known that the env proteins, which are the major proteins of the external envelope of retroviruses, have several properties that can be neuropathogenic. This has been shown for gp120 HIVenv, and, with respect to human endogenous retroviruses, only for the HERV-Wenv in multiple sclerosis, and HERV-Kenv in amyotrophic lateral sclerosis. In the latter two cases, transgenic mice containing the HERV-Wenv or HERV-Kenv genes developed a disease resembling, respectively, multiple sclerosis and amyotrophic lateral sclerosis.
It is a pity that Guo and colleagues, having obtained the transcriptome profiling of the samples, did not look for HERV-Fc1env transcripts, and, if positive for this, to HERV-Fc1 env protein staining of the lesions of the brain tissues. These studies are easily done, and would strengthen the relationships between AD’s tauopathy and HERV-Fc1 involvement in neurodegeneration.
References:
Borna H, Assadoulahei K, Riazi G, Harchegani AB, Shahriary A.
Structure, Function and Interactions of Tau: Particular Focus on Potential Drug Targets for the Treatment of Tauopathies.
CNS Neurol Disord Drug Targets. 2018;17(5):325-337.
PubMed.
Comments
Deceased
I first heard of the idea of transposon mobilization in human brain disease from Joshua Dubnau, then at Cold Spring Harbor. Josh and his colleagues published a paper providing evidence for activation of transposable elements in association with TDP pathology in ALS and FTLD (Li et al., 2012). Here, Guo et al. use a large RNA-seq dataset to show correlations between tau pathology and activation of transcription from transposable element loci in human brain, and in flies transgenic for human tau. This is an intriguing paper, reinforcing an idea which sounds less novel now than it did in 2012.
For a tauist, there is a "black box" to this story. What is it about tau that might trigger such a catastrophic response in a neuron? The human brain work correlates transposon activation with neurofibrillary tangle formation, but in Drosophila such aggregation of over expressed human tau is rare, and cell death usually occurs without tau filament formation. It is also far from clear how transcription from these loci will relate to cell death. Is this a sign of a sick cell - or a critical step in tau-mediated cell death? The complexity of the transcriptional activation will make this difficult to dissect.
View all comments by Peter DaviesStony Brook School of Medicine
Almost half of our DNA content consists of ancient virus like elements called retrotransposons (RTEs). RTEs are capable of copying themselves and then reinserting their new copies into our chromosomes. Our genomes invest heavily in keeping this half of our DNA quiescent because the wholesale activation of RTEs is incredibly damaging to the cell in a variety of ways.
There is now accumulating evidence to support the hypothesis that RTEs may contribute to the cellular toxicity that causes a variety of age-dependent neurodegenerative disorders, including amyotrophic lateral sclerosis, frontotemporal dementia, and macular degeneration. This study adds evidence that RTEs may be at the heart of AD.
It is important to recognize that although RTEs make up a vast fraction of our DNA, they have been almost completely ignored by human genetics (and by most geneticists studying animal models) until recently. The reason for this is that the standard computational algorithms for examining troves of genomic data begin by throwing out almost all the RTE sequences before the analysis really gets going. That has to stop because evidence is accumulating that RTEs may contribute to many age-related diseases, including cancer and neurodegeneration, and even normal aging. So it now becomes quite important to stop ignoring RTE sequences. By focusing on RTE sequences, these authors have, in fact, found strong evidence that many RTEs are highly expressed in AD brains and in a Drosophila model of tau pathology. It now becomes important to ask if such expression is a cause of or a consequence of the disease state.
View all comments by Josh DubnauUniversity of Queensland
This study investigated whether transposable elements (TEs) are differentially expressed in association with tau pathology in Alzheimer's disease (AD), using human post-mortem brain samples and a Drosophila model of AD. With regards to the human analysis, the cohort size (n=636) is impressive, and the observation of association with neurofibrillary tangles is intriguing. I also very much appreciated the circumspect discussion.
At the same time, I have a lot of questions. For example, it is not clear to me how big an absolute expression change was required for a given TE group to be elevated and considered significantly associated with a given tau pathology parameter in the linear regression statistics. 2%, 20% or 200%? As the authors note, the only autonomous human TE (L1Hs) was not at all associated with tangles, making it unclear how TE activity would contribute to genomic instability. The authors propose instead that some TE families are activated by chromatin relaxation promoted by tau pathology, which has been reported previously, and this makes more sense.
It is also interesting that the 5' end of one of the significant L1 subfamily hits (e.g. L1MB4_5 - copy number I think < 100) is significantly associated with tangles, whilst the 3' end of the same subfamily (L1MB4 - copy number ~9,200) is not. Why is that and how does it fit with the chromatin relaxation model? It would be useful to know the copy numbers of all of the human TE families with significant associations in the study.
Finally, I would be interested to see if these batch results replicate if one simply aligns the RNA-seq directly to the genome and intersects those data with the genomic coordinates of each individual TE; all except for the youngest TE subfamilies are highly mappable with the length and quality of RNA-seq reads available here, and this could generate some interesting examples of individual TE copies that are associated with pathology. In sum, the study is preliminary but very interesting and leads to many additional questions.
View all comments by Geoff FaulknerUniversity of Sassari
This paper by Guo and colleagues on the activation of transposable elements, and, in particular, of the LTR of HERV-Fc1 endogenous retrovirus in tau-associated lesions of human brains from patients who died with Alzheimer’s disease is interesting, as AD prevalence is increasing in elderly. Tau mechanisms of action were not specified in the paper, but a recent review notes that “although tau protein is well-known for its key role in stabilizing and organization of axonal microtubule network, it bears a broad range of functions including DNA protection and participation in signalling pathways,” and that tau can be modified by a variety of cellular enzymes, which in turn broaden tau’s function and interaction spectrum (Borna et al., 2018). This means that, in tauopathies, multiple mechanisms can be modified at different levels in DNA conformation and expression, and that it is hard to discriminate which effect is disease-related, and which is not.
In this study the only statistically significant alteration detected for transposable elements was in the transcriptional signature of HERV-Fc1 LTR. This could be relevant to correlate HERV-Fc1 to neurodegeneration, pending that transcripts driven by the LTR promoter are associated with some potentially neuropathogenic HERV-Fc1 ORFs, especially products of the env gene. It is known that the env proteins, which are the major proteins of the external envelope of retroviruses, have several properties that can be neuropathogenic. This has been shown for gp120 HIVenv, and, with respect to human endogenous retroviruses, only for the HERV-Wenv in multiple sclerosis, and HERV-Kenv in amyotrophic lateral sclerosis. In the latter two cases, transgenic mice containing the HERV-Wenv or HERV-Kenv genes developed a disease resembling, respectively, multiple sclerosis and amyotrophic lateral sclerosis.
It is a pity that Guo and colleagues, having obtained the transcriptome profiling of the samples, did not look for HERV-Fc1env transcripts, and, if positive for this, to HERV-Fc1 env protein staining of the lesions of the brain tissues. These studies are easily done, and would strengthen the relationships between AD’s tauopathy and HERV-Fc1 involvement in neurodegeneration.
References:
Borna H, Assadoulahei K, Riazi G, Harchegani AB, Shahriary A. Structure, Function and Interactions of Tau: Particular Focus on Potential Drug Targets for the Treatment of Tauopathies. CNS Neurol Disord Drug Targets. 2018;17(5):325-337. PubMed.
View all comments by Antonina DoleiMake a Comment
To make a comment you must login or register.